impact of vascular endothelial growth factor-a expression, thrombospondin-2 expression, and microvessel density on the treatment effect of bevacizumab in metastatic colorectal cancer Impact of Vascular Endothelial Growth Factor-A Expression, Thrombospondin-2 Expression, and Microvessel Density on the Treatment Effect of Bevacizumab in Metastatic Colorectal Cancer

Impact of Vascular Endothelial Growth Factor-A Expression, Thrombospondin-2 Expression, and Microvessel Density on the Treatment Effect of Bevacizumab in Metastatic Colorectal Cancer

  1. Hartmut Koeppen
  1. From the Departments of Pathology, Biostatistics, Bioinformatics, and BioOncology, Genentech Inc, South San Francisco; the University of California at Los Angeles, Los Angeles, CA; and the Department of Medical Oncology and Transplantation, Duke University Medical Center, Durham, NC
  1. Address reprint requests to Adrian Jubb, Academic Unit of Pathology, Algernon Firth Building, University of Leeds, Leeds, LS2 9JT, United Kingdom; e-mail: adrianjubb{at}gmail.com
 
Next Section

Abstract

Purpose Bevacizumab is a monoclonal antibody to vascular endothelial growth factor-A (VEGF). In the pivotal trial in metastatic colorectal cancer (mCRC), addition of bevacizumab to first-line irinotecan, fluorouracil, and leucovorin (IFL) significantly prolonged median survival. The aim of these retrospective subset analyses was to evaluate VEGF, thrombospondin-2 (THBS-2), and microvessel density (MVD) as prognostic factors and/or predictors of benefit from bevacizumab.

Patients and Methods In the pivotal trial, 813 patients with untreated mCRC were randomly assigned to receive IFL plus bevacizumab or placebo. Of 312 tissue samples collected (285 primaries, 27 metastases), outcome data were available for 278 (153 bevacizumab, 125 placebo). Epithelial and stromal VEGF expression were assessed by in situ hybridization (ISH) and immunohistochemistry on tissue microarrays and whole sections. Stromal THBS-2 expression was examined by ISH on tissue microarrays. MVD was quantified by Chalkley count. Overall survival was associated with these variables in retrospective subset analyses.

Conclusion These exploratory analyses suggest that in patients with mCRC addition of bevacizumab to IFL improves survival regardless of the level of VEGF or THBS-2 expression, or MVD.

Previous SectionNext Section

INTRODUCTION

Vascular endothelial growth factor-A (VEGF) is a secreted ligand with specific receptors that are expressed principally by angioblasts and endothelial cells.1 The interaction of VEGF with its receptors plays a major role in regulating both physiologic and pathologic angiogenesis, and a wide range of human malignancies upregulate their expression of VEGF to varying degrees during tumorigenesis.2 In preclinical studies a murine antihuman monoclonal antibody (mAb) directed against VEGF was shown to inhibit the growth of human tumor xenografts.3 This finding prompted a clinical trial program to assess the efficacy of a humanized variant of this antibody (bevacizumab) as a treatment for various cancers.

Colorectal cancer (CRC) is the second most common cause of cancer deaths worldwide, with 945,000 new cases and 492,000 CRC-related deaths in 2000.4 In patients with metastatic CRC (mCRC), the addition of bevacizumab to an irinotecan, fluorouracil, and leucovorin (IFL) regimen was shown recently to significantly prolong the median duration of survival (from 15.6 to 20.3 months, corresponding to a hazard ratio (HR) for death of 0.66; P < .001).5

While this result demonstrates the potential of antiangiogenic therapy, the identification of biomarkers that may influence response to such treatment is of considerable interest. High levels of tumor VEGF expression, increased circulating VEGF, and increased microvessel density (MVD) have each been associated with poorer survival and an increased incidence of disease recurrence and distant metastasis in adenocarcinomas of the colon and rectum.6-21 As clinical responsiveness to some targeted therapies (for example, trastuzumab in breast cancer) is strongly influenced by the level of expression of the target,22 it would seem reasonable to evaluate tumor VEGF expression as a predictor of responsiveness to bevacizumab therapy.

Endogenous negative regulators of angiogenesis may also play a prominent role in tumor behavior. For instance, members of the thrombospondin (THBS) gene family are reported to have antiangiogenic properties, (eg, inhibition of VEGF, matrix metalloproteinases, and basic fibroblast growth factor)23-26 and both preclinical and clinical studies suggest that they can inhibit tumor growth.27-36 Both THBS-1 and THBS-2 are expressed frequently by stromal cells in the tumor microenvironment and are reported to have prognostic significance in several malignancies.37-38 Expression of THBS-2 in CRC is associated with the inhibition of angiogenesis and a reduced frequency of distant metastasis,33 and has been implicated in the inhibition of angiogenesis even in the presence of high tumor-cell VEGF expression.31 Thus, there is a biologic rationale to suggest that host factors involved in the negative regulation of angiogenesis may serve as predictors of response to antiangiogenic therapies.

In addressing these hypotheses, we conducted exploratory, retrospective subset analyses to evaluate VEGF expression, THBS-2 expression, and MVD as prognostic factors and/or predictors of benefit from bevacizumab in mCRC.

Previous SectionNext Section

PATIENTS AND METHODS

Patients and Study Design

Patient recruitment and trial design have been described elsewhere.5 In brief, previously untreated patients with mCRC were randomly assigned to receive IFL with placebo, IFL with bevacizumab, or fluorouracil and leucovorin with bevacizumab. At the time of enrollment, informed consent was obtained from trial participants to permit research on their archived tissue. The survival analyses described herein included tumor samples from only patients in the treatment arms that received irinotecan (placebo plus IFL [n = 411] or bevacizumab plus IFL [n = 402]). All experiments and analyses were performed blinded to clinical outcome.

Tissue Samples and Tissue Microarray Construction

Formalin-fixed paraffin-embedded CRC tissue blocks and corresponding pathology reports were obtained for 312 patients from multiple centers. The case series comprised 27 metastases (one peritoneal, five omental, two lymph node, three small intestine, three lung, 11 liver, and two ovarian) and 285 primary tumors. Tissue microarrays (TMAs) were assembled as described previously.39 Immunohistochemistry (IHC) and in situ hybridization (ISH) were performed on TMAs (297 patients) and/or on whole sections (235 patients).

In Situ Hybridization

cDNA probe templates were generated from human kidney Marathon-ready cDNA (BD Clontech, Palo Alto, CA). Primers were designed to amplify a 604–base pair (bp) fragment of VEGF mRNA (Genbank accession M32977), 439 bps of THBS-2 mRNA (NM_003247), and a 291 bp fragment of β-actin mRNA (NM_001101). VEGF ISH probes did not discriminate between mature mRNA species encoding the different VEGF isoforms. Sense and antisense primers included T7 and T3 RNA polymerase initiation sites, respectively (VEGF sense: 5′-T7-GGGCCTCCGAAACCATGAACT-3′, VEGF antisense: 5′-T3-TCCTCCTGCCCGGCTCAC-3′, THBS-2 sense: 5′-T7-CTACATCTCCAACGCCAACCA-3′, THBS-2 antisense: 5′-T3-GTCGTCGTCCCGGTCAGT-3′, β-actin sense: 5′-T7-GCTGCCTGACGGCCAGGTC-3′, β-actin antisense: 5′-T3-GAGTACTTGCGCTCAGAGGAG-3′). Riboprobe synthesis, hybridization, developing, and analysis were carried out as described previously.2,40-42

Hybridization of antisense β-actin riboprobes was confirmed in all tissues before biomarker analysis. Whole sections of healthy human kidney and human tumor cell pellets were included in each experiment as controls for appropriate hybridization of antisense VEGF riboprobes (Figs 1 and 2). Sense VEGF and THBS-2 riboprobes were employed as negative controls for the specificity of hybridization. TMA cores were scored semiquantitatively on a scale of zero (no expression) to three (very strong signal), according to the overall intensity of the hybridization signal in the epithelium (VEGF) or stroma (THBS-2). VEGF expression in whole sections was scored on an identical scale to reflect the greatest signal intensity observed from at least 10% of neoplastic cells (VEGF maximum), the overall signal intensity observed from greater than 50% of neoplastic cells (VEGF overall), and the signal intensity observed from the stromal cells (VEGF stroma). TMA scoring was determined by a consensus of two pathologists (K.J.H. and H.K.) on a double-headed microscope; whole sections were scored by a single pathologist (H.K.). The hybridization signal intensities in the cell pellets were employed as reference points to assign VEGF scores in each experimental run; 786-0 cell pellets were scored three, and MCF7 cell pellets scored one by ISH (Figs 1 and 2). All ISH on TMAs, including all THBS-2 analyses, was performed in a single experiment. Representative VEGF and THBS-2 scores are illustrated in Figures 3 and 4.

Immunohistochemistry

IHC was performed on freshly-cut tissue sections as described previously.2,41 Sections were incubated for 1 hour at room temperature with a primary antihuman VEGF mAb at 1 μg/mL (clone 26503, recognizing the 121 and 165 amino acid isoforms of VEGF; R&D Systems, Minneapolis, MN) or with a primary antihuman CD34 mAb at 2 μg/mL (clone QBEnd/10; NovoCastra, Newcastle on Tyne, United Kingdom). Negative control slides were incubated with an isotype-matched mouse immunoglobulin culture supernatant (Dako Cytomation, Copenhagen, Denmark) at 1 or 2 μg/mL, as appropriate, in place of the primary antibody. Whole sections of healthy human kidney and human tumor cell pellets were included in each experiment as controls for appropriate VEGF and CD34 immunostaining (Figs 1 and 2). To assess VEGF expression, TMA cores and whole sections were scored semiquantitatively on a scale of zero (no expression) to three (very strong signal), according to the intensity of chromogen deposition in the majority of neoplastic cells. TMAs and whole sections were scored by a single pathologist (H.K.). The hybridization signal intensities in the cell pellets were employed as reference points to assign VEGF scores in each experimental run; 786-0 cell pellets were scored three, and MCF7 cell pellets scored zero by IHC (Figs 1 and 2). All IHC on TMAs was performed in a single experiment. Representative scores are illustrated in Figure 5.

MVD Scoring

CD34-stained whole sections were scanned into tiled, digital color images at 0.64-μ pixel size using a custom-automated image acquisition application for a Nikon E1000 microscope (Technical Instruments, Burlingame, CA). Images were processed by a custom software algorithm that automates the Chalkley counting procedure described previously.43,44 The algorithm exhaustively evaluated Chalkley scores over the image by varying the position and rotation of a virtual Chalkley field (equivalent to a 20× high-power objective with a 10× eyepiece magnification). The virtual Chalkley field was rotated through 360° in 1° increments at a rectangular array of positions spaced one Chalkley field radius apart and covering the entire image. The coincidence of a Chalkley point with an immunohistochemically-stained pixel was detected by colorimetric filtering. The three top-scoring locations, separated by at least one Chalkley field radius, were reported along with other statistics and diagnostic images. Images were reviewed for accuracy, and the mean vessel counts per Chalkley field were recorded. (Details of the algorithm are available from the authors by request.)

Statistical Analysis

The prognostic and predictive value of VEGF, THBS-2, and MVD were summarized using survival analysis methods. High expression of VEGF and THBS-2 by ISH was defined as score > 2. High expression of VEGF by IHC was defined as score > 1. The highest level of each factor was chosen as the high category unless there were too few patients in the group for purposes of making comparisons. MVD was assessed as a continuous variable. Due to sample considerations, the statistical significance of differences between subgroups were assessed using the log-rank test without adjustment for stratification factors, such as location of the primary tumor, the number of organ sites with metastases or Eastern Cooperative Oncology Group (ECOG) performance status. Median survival times were estimated from Kaplan-Meier curves. HRs relative to treatment within each marker category were determined using a Cox model with only a single term for treatment included in the model. HRs relative to the marker within each treatment group were determined from a Cox model with a single term for marker included in the model. 95% CIs for the HRs were constructed on the basis of Wald-type tests. Interactions between the marker and bevacizumab treatment were assessed using a Cox model with treatment, marker, and the treatment-marker interaction terms in the model. The proportional hazards assumption was assessed for the study as a whole by fitting a Cox model with a term for treatment and a time-dependent term that was zero for time less than t and treatment for time greater than t. We let t assume a number of values and found that the proportional hazards assumptions are not contradicted. There were too few patients to evaluate the proportional hazards assumption within each category of the biomarkers. P values < .05 were considered significant, whereas only P values < .10 were considered trends. SAS software (SAS Institute, Cary, NC) was used to carry out the statistical analyses.

Previous SectionNext Section

RESULTS

Study Group Characteristics and Frequency Data

The cohort of patients in this study had demographic and pathologic characteristics that were representative of the pivotal trial population (Table 1). At the time of analysis, outcome data were available for 278 of the 312 patients examined (34% of pivotal trial participants). The remaining 34 of the 312 patients were treated with fluorouracil, leucovorin, and bevacizumab (arm 3 of the pivotal trial) and were not included in these analyses. The survival benefit from the addition of bevacizumab was not significantly different in the subset of 278 patients compared with the entire study population (HRs for bevacizumab plus IFL v placebo plus IFL = 0.57 [95% CI, 0.43 to 0.76] in the subset compared with 0.66 [95% CI, 0.54 to 0.81] in the entire study population; Table 1). Analyses of tumors from the 278 patients yielded informative data on 169 to 226 patients (Supplementary Fig 1, online only). Results were not available for the remaining cases because of limited amounts of tissue or technical assay failures.

Hybridization of the antisense VEGF riboprobe was more frequently observed over neoplastic cells than stromal cells, or healthy epithelium (Figs 1, 2, and 3). Considerable intratumor heterogeneity was noted in whole sections, although expression at the invasive edge of the cancer was not appreciably different from the remainder (data not shown). Less frequently, silver grains (indicative of mRNA expression) were seen over tumor-associated stromal cells (Fig 4). Signal intensity over neoplastic cells was often greatest in areas adjacent to tumor necrosis and declined with increasing distance from the necrotic focus (Fig 4). In certain cases, the inflammatory cell component of tumor-associated inflammatory infiltrates was found to express VEGF (Fig 4). Healthy crypts adjacent to the tumor were occasionally observed to express VEGF, in most cases at levels below the lowest-expressing tumor cells. Hybridization of the antisense THBS-2 riboprobe was most frequently observed covering tumor-associated stromal cells (Fig 3); THBS-2 message was absent from adjacent, healthy colon. Sense riboprobes did not demonstrate appreciable hybridization greater than background.

Immunolabeling of VEGF demonstrated strong chromogen deposition over healthy kidney glomeruli in each run of experiments, similar to the ISH signal. IHC for VEGF in CRC specimens was most intense in the cytoplasm of neoplastic cells, where a granular pattern of staining could be observed (Fig 5). No appreciable staining was seen in adjacent healthy crypts. Weak immunoreactivity was observed frequently in the extracellular matrix and stromal cells, including ganglion cells of the myenteric plexus (Fig 5). Endothelium showed evidence of VEGF staining in both tumor-associated stroma and stroma found in healthy colonic mucosa adjacent cancer in the same tissue block (Fig 5). Tissue sections treated with isotype control antibodies did not exhibit appreciable staining.

Association of Biomarkers With Treatment Outcome

In all biomarker subgroups, point estimates of HRs for death were less than one for patients treated with bevacizumab plus IFL when compared with those treated with placebo plus IFL (Fig 6A). None of the methods employed to assess VEGF expression discriminated between subjects more or less likely to benefit from the addition of bevacizumab to first-line IFL. Patients whose tumors had a high THBS-2 score (> 2) showed an improved median survival with bevacizumab treatment (HR = 0.11; 95% CI, 0.02 to 0.51) when compared to patients with a low THBS-2 score (HR = 0.65; 95% CI, 0.41 to 1.02). This observation was not statistically significant by interaction analysis (P = .2228; Table 2). Quantitation of VEGF and THBS-2 expression by phosphor-image analysis of ISH signal intensities did not confer additional value to the assessment of these biomarkers (data not shown). Analogous results were found in analyses of progression-free survival and objective response rate (data not shown).

Association of Biomarkers With Prognosis

VEGF and THBS-2 expression were not significant prognostic factors for duration of survival, irrespective of the treatment received (Fig 6B). Comparable results were found in analyses of progression-free survival and objective response rate (data not shown).

Association of MVD With Treatment Outcome and Prognosis

MVD was scored on 235 patients, of which 209 patients were included in the IFL-containing arms of the trial, 94 patients received IFL plus placebo and 115 patients received IFL plus bevacizumab. Distributions of MVD scores were not different between placebo (median, 7.8; interquartile range, 6.3 to 9.3) and bevacizumab treatment groups (median, 8.0; interquartile range, 6.3 to 9.3). When MVD was assessed as a continuous variable, it was not observed to be a significant prognostic factor (HR = 0.963; 95% CI, 0.899 to 1.032), or a significant predictive marker (HR = 1.002; 95% CI, 0.872 to 1.152) in terms of overall survival. These findings were not altered when varying cutoffs were chosen to define high and low MVD (data not shown). Comparable results were found in analyses of progression-free survival and objective response rate (data not shown). However, the MVD score was significantly associated with the maximum epithelial (Kruskal-Wallis analysis of variance [ANOVA] statistic = 8.58; P = .01), overall epithelial (Kruskal-Wallis ANOVA statistic = 22.86; P < .0001), and stromal (Kruskal-Wallis ANOVA statistic = 8.37; P = .04) VEGF ISH scores in whole sections.

Previous SectionNext Section

DISCUSSION

There is a biologic rationale to suggest that the level of VEGF expression by a tumor would determine its responsiveness to bevacizumab.6,7,10-13,45-47 Bevacizumab binds directly to VEGF, inhibiting its biologic activity3,48,49 and leaving the vasculature unable to support progressive tumor growth.3,50,51 Nevertheless, the examination of VEGF expression in this series did not reveal patient subgroups with differential responses to bevacizumab therapy (Fig 6A). Indeed, all CRCs in this series exhibited some level of VEGF expression in certain cell populations, and it may be that the distinction between high versus low VEGF expression is not clinically important. While these findings are of considerable interest, all analyses were performed retrospectively on relatively small subsets of predominantly primary tumors. Interpretation of the data must be performed with care, and these findings warrant confirmation in a larger series.

In an initial survey, VEGF expression was examined in TMAs, permitting experimental standardization and the quantitation of ISH signal intensities. In this arrayed series, the level of VEGF expression was not a prognostic factor, nor a predictive factor of benefit from bevacizumab treatment (Fig 6). However, scoring VEGF expression by tumor region and cell type has also been reported previously to have prognostic significance.8,9 Therefore, a more detailed analysis of VEGF expression was conducted on whole sections. Nevertheless, in these analyses, both epithelial and stromal VEGF expression were independent of patient prognosis and response to bevacizumab in terms of overall survival (Fig 6). This is in keeping with the literature, which contains many conflicting reports on the prognostic significance of VEGF expression in CRC.6-13,47,52,53 Discrepancies within the literature most likely are a result of clinicopathologic differences between cohorts, and methodologic diversity. Specifically, our protocol stipulated collection of archived tissue specimens, the vast majority of which are primary tumor tissue and may not accurately reflect the biomarker status of metastatic disease. VEGF expression is reported to be relatively higher in colorectal metastases2 compared with the primary disease, and may also vary with the location of the metastasis.54,55 However, Kuramochi et al56 recently reported no significant difference in the level of VEGF mRNA expression in a series of matched primary CRCs and liver metastases. Until this issue is resolved, in future clinical trials it will be important to obtain metastatic biopsies prospectively to allow the assessment of biomarker status in the treated disease.

In addition to their synthesis of proangiogenic molecules, tumor-associated stromal cells also express antiangiogenic factors, including members of the THBS gene family. Both THBS-1 and THBS-2 are expressed in the stroma of CRC,33 although analyses have conferred a greater clinical significance to THBS-2 expression.33,37,38,57 One may conjecture that tumors with a high level of THBS expression are more dependent on the expression of proangiogenic molecules to tip the balance in favor of angiogenesis. In this environment, the impact of bevacizumab might be bolstered by the antiangiogenic effects of THBS-2.23-26 These observations, in addition to the availability of a robust THBS-2 assay, led us to investigate the prognostic and predictive impact of THBS-2 in our substudy population. Although THBS-2 was not a significant prognostic factor (Fig 6B), the overall survival benefit from the addition of bevacizumab was greater in the subgroup that scored high for THBS-2 expression (Fig 6A). However, this observation did not reach statistical significance in an interaction analysis (Table 2). Therefore, these data do not have implications for clinical practice, as all patient subgroups benefited from the addition of bevacizumab to first-line IFL. In addition, the limitations of this retrospective, subset analysis must be addressed prospectively in a larger population before any definitive conclusions can be drawn.

MVD is believed to summarize the effects of all angiogenic regulators, and therefore may better predict patient outcome than the analysis of a single growth factor or signal-transduction pathway. Assessments of MVD are frequently employed as surrogate measures of angiogenesis in CRC7,8,45-47,53,58 and there are numerous reports describing associations of MVD with both VEGF expression and prognosis.7,45,47 In this series, MVD was not a significant prognostic factor or a predictor of benefit from bevacizumab treatment. This discrepancy may reflect differences in the clinical behavior of primary malignancies and metastatic disease. Angiogenesis is a reactive phenomenon, and not necessarily a stable phenotype that is comparable in primary cancers and their metastases. Moreover, the prognostic impact of MVD may be reduced in more advanced malignancies. Scoring the fraction of immature vessels, which are most susceptible to anti-VEGF therapy,51,59 might provide additional information. At present, though, the significance of MVD in assessing response to antiangiogenic therapy remains unclear.

Human mCRCs treated with bevacizumab eventually progress.5,50 Although we cannot exclude the possibility that a different dosing regimen may be more efficacious, response may be best predicted by screening for the acquisition of a phenotype that permits escape from VEGF inhibition.50 This could be achieved by expression profiling60-62 or analysis of tumor physiology51,63 before and after treatment with bevacizumab. Specific tumor types demonstrate coexpression of VEGF and other angiogenic factors,64,65 raising the possibility that complete inhibition of angiogenesis might require the blockade of more than one signaling pathway.66,67 For instance, it has been suggested that predominantly immature endothelium, devoid of pericytes, regresses on treatment with bevacizumab.51,67,68 Endothelial survival signals produced by neoplastic cells,65 tumor stroma,69 and/or pericytes68 are believed to confer resistance to anti-VEGF treatment. Examples include angiopoietin-1/tie-2 and platelet-derived growth factor signaling.67,68,70 Therefore, examining the expression of genes involved in pericyte recruitment and pericyte-endothelial paracrine signaling may yield biomarkers that predict escape from bevacizumab and new strategies for combination antiangiogenic therapy.

These exploratory analyses suggest that, in patients with mCRC, addition of bevacizumab to first-line IFL improves survival regardless of the level of VEGF or THBS-2 expression, or MVD. More fundamental research on the appropriate clinical material is required to fully understand the complexity of angiogenesis and patient response to antiangiogenic therapy.

Previous SectionNext Section

Authors’ Disclosures of Potential Conflicts of Interest

Although all authors completed the disclosure declaration, the following author or immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed discription of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.

Authors Employment Leadership Consultant Stock Honoraria Research Funds Testimony Other
Adrian M. Jubb Genentech Inc (NR)
Herbert I. Hurwitz Genentech Inc (B)
Wei Bai Genentech Inc (NR) Genentech Inc (B)
Eric B. Holmgren Genentech Inc (NR) Genentech Inc (B)
Patti Tobin Genentech Inc (NR) Genentech Inc (B)
A. Steven Guerrero Genentech Inc (NR) Genentech Inc (B)
Fairooz Kabbinavar Genentech Inc (A)
Scott N. Holden Genentech Inc (NR) Genentech Inc (B)
William F. Novotny Genentech Inc (NR) Genentech Inc (B)
Gretchen D. Frantz Genentech Inc (NR) Genentech Inc (B)
Kenneth J. Hillan Genentech Inc (NR) Genentech Inc (B)
Hartmut Koeppen Genentech Inc (NR) Genentech Inc (B)

  • Dollar Amount Codes (A) < $10,000 (B) $10,000–99,000 (C) ≥ $100,000 (N/R) Not Required

    • Previous SectionNext Section

    Conception and design: Adrian M. Jubb, Kenneth J. Hillan, Hartmut Koeppen

    Provision of study materials or patients: Herbert I. Hurwitz, Fairooz Kabbinavar, William F. Novotny, Kenneth J. Hillan

    Collection and assembly of data: Adrian M. Jubb, Wei Bai, Patti Tobin, A. Steven Guerrero, Gretchen D. Frantz, Kenneth J. Hillan, Hartmut Koeppen

    Data analysis and interpretation: Adrian M. Jubb, Eric B. Holmgren, Hartmut Koeppen

    Manuscript writing: Adrian M. Jubb, A. Steven Guerrero, Scott N. Holden

    Final approval of manuscript: Hartmut Koeppen

    Previous SectionNext Section

    GLOSSARY

    Angiogenesis:
    The process involved in the generation of new blood vessels. While this is a normal process that naturally occurs and is controlled by “on” and “of” switches, blocking tumor angiogenesis (antiangiogenesis) disrupts the blood supply to tumors, thereby preventing tumor growth.
    Microvessel density (Chalkley count):
    A microvessel density scoring system whereby tissue sections stained for tumor vasculature are processed by an observer or a custom software algorithm. The real or virtual Chalkley graticule is positioned over each of three vascular hotspots, determined at low power. The greatest number of vessels that coincide with the 25 points on the Chalkley graticule (after it is rotated through 360° at 20× objective and 10× eyepiece magnification) is averaged for the three hotspot areas to define the Chalkley score.
    Monoclonal antibody:
    An antibody that is secreted from a single clone of an antibody-forming cell. Large quantities of monoclonal antibodies are produced from hybridomas, which are produced by fusing single antibody-forming cells to tumor cells. The process is initiated when a mouse is immunized initially against a particular antigen, stimulating the production of antibodies targeted to different epitopes of the antigen. Antibody-forming cells are subsequently isolated from the spleen. By fusing each antibody-forming cell to tumor cells, hybridomas can be generated each with a different specificity and targeted against a different epitope of the antigen.
    Pericytes:
    Fibroblastic/smooth muscle-like cells found in close contact with endothelial cells in small blood vessels and capillaries, where they function as regulators of blood vessel formation and function, in particular contributing to vascular integrity.
    THBS (thrombospondin):
    A family of extracellular adhesive proteins with five members (THBS-1 through THBS-4, and cartilage oligomeric matrix protein [COMP]), THBSs play a role in several cellular processes, including platelet aggregation and angiogenesis.
    Tissue microarray:
    Used to analyze the expression of genes of interest simultaneously in multiple tissue samples, tissue microarrays consist of hundreds of individual tissue samples placed on slides ranging from 2 to 3 mm in diameter. Using conventional histochemical and molecular detection techniques, tissue microarrays are powerful tools to evaluate the expression of genes of interest in tissue samples. In cancer research, tissue microarrays are used to analyze the frequency of a molecular alteration in different tumor type, to evaluate prognostic markers, and to test potential diagnostic markers.
    VEGF (vascular endothelial growth factor):
    VEGF is a cytokine that mediates numerous functions of endothelial cells including proliferation, migration, invasion, survival, and permeability. VEGF is also known as vascular permeability factor. VEGF naturally occurs as a glycoprotein and is critical for angiogenesis. Many tumors overexpress VEGF, which correlates to poor prognosis. VEGF-A, -B, -C, -D, and -E are members of the larger family of VEGF-related proteins.

    Previous SectionNext Section

    Fig 1.
    View larger version:
    Fig 1.

    Cell pellet controls for vascular endothelial growth factor-A (VEGF) expression and localization by in situ hybridization and immunohistochemistry. (A) In situ hybridization, bright field, hematoxylin and eosin; (B) dark field, VEGF antisense; (C) dark field, VEGF sense; (D) dark field, β-actin antisense; and (E) immunohistochemistry, bright field, VEGF.

    Fig 2.
    View larger version:
    Fig 2.

    Healthy human kidney controls for vascular endothelial growth factor-A (VEGF) expression and localization by in situ hybridization and immunohistochemistry. The arrow indicates the renal glomerulus. (A) In situ hybridization, bright field, hematoxylin and eosin; (B) dark field, VEGF antisense; and (C) immunohistochemistry, bright field, VEGF.

    Fig 3.
    View larger version:
    Fig 3.

    Representative tumor cores illustrating the in situ hybridization scoring system for vascular endothelial growth factor-A (VEGF) and thrombospondin-2 (THBS-2) expression. β-actin was employed as a control for RNA integrity, and sense riboprobes were used to assess the specificity of hybridization. VEGF expression was observed predominantly in the neoplastic cells, in contrast to THBS-2 expression, which localized to the stroma. (A) In situ hybridization, bright field, hematoxylin and eosin; (B) dark field, VEGF antisense; (C) dark field, VEGF sense; (D) dark field, THBS-2 antisense; (E) dark field, THBS-2 sense; and (F) dark field, β-actin antisense.

    Fig 4.
    View larger version:
    Fig 4.

    Bright field (A-D) and dark field (E-H) vascular endothelial growth factor-A (VEGF) in situ hybridization images. Stromal expression (red arrows) was observed in tumors with (A, E) and without (B, F) expression by the neoplastic cells (black/white arrows). High levels of VEGF adjacent regions of necrosis (black/white to red arrows, C, G). Hybridization was also observed over inflammatory infiltrates (black/white arrow, D, H).

    Fig 5.
    View larger version:
    Fig 5.

    Representative tumor cores illustrating the immunohistochemistry scoring system for vascular endothelial growth factor-A (VEGF) expression. VEGF protein was predominantly seen in the cytoplasm of neoplastic cells, in addition to labeling of endothelium and ganglion cells. (A) Score 0; (B) score 1; (C) score 2; (D) score 3; (E) endothelium (arrow); and (F) ganglion cell (arrow).

    Fig 1A.
    View larger version:
    Fig 1A.

    Attrition of patient numbers from the total 312 acquired to the subsets of 169 to 226 patients on whom analyses were performed. TMA, tissue microarray; VEGF, vascular endothelial growth factor-A; ISH, in situ hybridization; THBS-2, thrombospondin-2; IHC, immunohistochemistry; MVD, microvessel density; BV, bevacizumab; IFL, irinotecan, fluorouracil, and leucovorin.

    View this table:
    Table 1.

    Demographic and Clinical Characteristics of Study Populations

    View this table:
    Table 2.

    Interaction Analyses Between Each Biomarker and Treatment Subgroup

    Previous SectionNext Section

    Acknowledgments

    We thank all patients and investigators who participated in the phase III clinical trial of bevacizumab in metastatic colorectal cancer. Members of Genentech’s anatomical pathology and oligonucleotide synthesis groups were instrumental in the implementation of this study.

    Previous SectionNext Section

    Footnotes

    • Presented in part at the American Society of Clinical Oncology Gastrointestinal Cancers Symposium, Miami, FL, January 27-29, 2005.

      Terms in blue are defined in the glossary, found at the end of this article and online at www.jco.org.

      Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article.

    • Received February 18, 2005.
    • Accepted July 29, 2005.
    Previous Section
     

    REFERENCES

    1. ↵
      Ferrara N, Davis-Smyth T: The biology of vascular endothelial growth factor. Endocr Rev 18:4-25, 1997
      Abstract/FREE Full Text
    2. ↵
    3. ↵
      Kim KJ, Li B, Winer J, et al: Inhibition of vascular endothelial growth factor-induced angiogenesis suppresses tumour growth in vivo. Nature 362:841-844, 1993
      CrossRefMedline
    4. ↵
      Parkin DM: Global cancer statistics in the year 2000. Lancet Oncol 2:533-543, 2001
      CrossRefMedline
    5. ↵
      Hurwitz H, Fehrenbacher L, Novotny W, et al: Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 350:2335-2342, 2004
      CrossRefMedline
    6. ↵
      Cascinu S, Graziano F, Catalano V, et al: An analysis of p53, BAX and vascular endothelial growth factor expression in node-positive rectal cancer: Relationships with tumour recurrence and event-free survival of patients treated with adjuvant chemoradiation. Br J Cancer 86:744-749, 2002
      CrossRefMedline
    7. ↵
      Galizia G, Lieto E, Ferraraccio F, et al: Determination of molecular marker expression can predict clinical outcome in colon carcinomas. Clin Cancer Res 10:3490-3499, 2004
      Abstract/FREE Full Text
    8. ↵
      Kaio E, Tanaka S, Kitadai Y, et al: Clinical significance of angiogenic factor expression at the deepest invasive site of advanced colorectal carcinoma. Oncology 64:61-73, 2003
      CrossRefMedline
    9. ↵
      Khorana AA, Ryan CK, Cox C, et al: Vascular endothelial growth factor, CD68, and epidermal growth factor receptor expression and survival in patients with stage II and stage III colon carcinoma: A role for the host response in prognosis. Cancer 97:960-968, 2003
      CrossRefMedline
    10. Konno H, Abe J, Kaneko T, et al: Urokinase receptor and vascular endothelial growth factor are synergistically associated with the liver metastasis of colorectal cancer. Jpn J Cancer Res 92:516-523, 2001
      CrossRef
    12. Lee JC, Chow NH, Wang ST, et al: Prognostic value of vascular endothelial growth factor expression in colorectal cancer patients. Eur J Cancer 36:748-753, 2000
      CrossRefMedline
    13. ↵
      Maeda K, Nishiguchi Y, Yashiro M, et al: Expression of vascular endothelial growth factor and thrombospondin-1 in colorectal carcinoma. Int J Mol Med 5:373-378, 2000
      Medline
    14. Davies MM, Jonas SK, Kaur S, et al: Plasma vascular endothelial but not fibroblast growth factor levels correlate with colorectal liver mestastasis vascularity and volume. Br J Cancer 82:1004-1008, 2000
      CrossRefMedline
    15. Hyodo I, Doi T, Endo H, et al: Clinical significance of plasma vascular endothelial growth factor in gastrointestinal cancer. Eur J Cancer 34:2041-2045, 1998
      CrossRefMedline
    16. Kumar H, Heer K, Lee PW, et al: Preoperative serum vascular endothelial growth factor can predict stage in colorectal cancer. Clin Cancer Res 4:1279-2185, 1998
      Abstract
    17. Lissoni P, Rovelli F, Malugani F, et al: Changes in circulating VEGF levels in relation to clinical response during chemotherapy for metastatic cancer. Int J Biol Markers 18:152-155, 2003
      Medline
    18. Nakayama Y, Sako T, Shibao K, et al: Prognostic value of plasma vascular endothelial growth factor in patients with colorectal cancer. Anticancer Res 22:2437-2442, 2002
      Medline
    19. Takeda A, Shimada H, Imaseki H, et al: Clinical significance of serum vascular endothelial growth factor in colorectal cancer patients: Correlation with clinicopathological factors and tumor markers. Oncol Rep 7:333-338, 2000
      Medline
    20. Werther K, Sorensen S, Christensen IJ, et al: Circulating vascular endothelial growth factor six months after primary surgery as a prognostic marker in patients with colorectal cancer. Acta Oncol 42:837-845, 2003
      CrossRefMedline
    21. ↵
      Werther K, Christensen IJ, Nielsen HJ: Prognostic impact of matched preoperative plasma and serum VEGF in patients with primary colorectal carcinoma. Br J Cancer 86:417-423, 2002
      CrossRefMedline
    22. ↵
      Slamon DJ, Leyland-Jones B, Shak S, et al: Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med 344:783-792, 2001
      CrossRefMedline
    23. ↵
      Adams JC: Thrombospondins: Multifunctional regulators of cell interactions. Annu Rev Cell Dev Biol 17:25-51, 2001
      CrossRefMedline
    24. Bein K, Simons M: Thrombospondin type 1 repeats interact with matrix metalloproteinase 2: Regulation of metalloproteinase activity. J Biol Chem 275:32167-32173, 2000
      Abstract/FREE Full Text
    25. Iruela-Arispe ML, Lombardo M, Krutzsch HC, et al: Inhibition of angiogenesis by thrombospondin-1 is mediated by 2 independent regions within the type 1 repeats. Circulation 100:1423-1431, 1999
      Abstract/FREE Full Text
    26. ↵
      Simantov R, Febbraio M, Bornstein P, et al: The anti-angiogenic activity of thrombospondin-2 (TSP-2) is mediated by CD36: Evidence that the CD36 anti-angiogenic response may be activated by a family of thrombospondin type I repeat (TSR) domain proteins. Blood 98:822A, 2001 (abstr 3415)
    27. ↵
      Allegrini G, Goulette FA, Darnowski JW, et al: Thrombospondin-1 plus irinotecan: A novel antiangiogenic-chemotherapeutic combination that inhibits the growth of advanced human colon tumor xenografts in mice. Cancer Chemother Pharmacol 53:261-266, 2004
      CrossRefMedline
    28. Bocci G, Francia G, Man S, et al: Thrombospondin 1, a mediator of the antiangiogenic effects of low-dose metronomic chemotherapy. Proc Natl Acad Sci U S A 100:12917-12922, 2003
      Abstract/FREE Full Text
    29. Kyriakides TR, Leach KJ, Hoffman AS, et al: Mice that lack the angiogenesis inhibitor, thrombospondin 2, mount an altered foreign body reaction characterized by increased vascularity. Proc Natl Acad Sci U S A 96:4449-44954, 1999
      Abstract/FREE Full Text
    30. Rodriguez-Manzaneque JC, Lane TF, Ortega MA, et al: Thrombospondin-1 suppresses spontaneous tumor growth and inhibits activation of matrix metalloproteinase-9 and mobilization of vascular endothelial growth factor. Proc Natl Acad Sci U S A 98:12485-12490, 2001
      Abstract/FREE Full Text
    31. ↵
      Streit M, Riccardi L, Velasco P, et al: Thrombospondin-2: A potent endogenous inhibitor of tumor growth and angiogenesis. Proc Natl Acad Sci U S A 96:14888-14893, 1999
      Abstract/FREE Full Text
    32. Streit M, Stephen AE, Hawighorst T, et al: Systemic inhibition of tumor growth and angiogenesis by thrombospondin-2 using cell-based antiangiogenic gene therapy. Cancer Res 62:2004-2012, 2002
      Abstract/FREE Full Text
    33. ↵
      Tokunaga T, Nakamura M, Oshika Y, et al: Thrombospondin 2 expression is correlated with inhibition of angiogenesis and metastasis of colon cancer. Br J Cancer 79:354-359, 1999
      CrossRefMedline
    34. Tomii Y, Kamochi J, Yamazaki H, et al: Human thrombospondin 2 inhibits proliferation of microvascular endothelial cells. Int J Oncol 20:339-342, 2002
      Medline
    35. Volpert OV, Tolsma SS, Pellerin S, et al: Inhibition of angiogenesis by thrombospondin-2. Biochem Biophys Res Commun 217:326-332, 1995
      CrossRefMedline
    36. ↵
      Werther K, Christensen IJ, Brunner N, et al: Soluble vascular endothelial growth factor levels in patients with primary colorectal carcinoma: The Danish RANX05 Colorectal Cancer Study Group. Eur J Surg Oncol 26:657-662, 2000
      CrossRefMedline
    37. ↵
      Oshika Y, Masuda K, Tokunaga T, et al: Thrombospondin 2 gene expression is correlated with decreased vascularity in non-small cell lung cancer. Clin Cancer Res 4:1785-1788, 1998
      Abstract
    38. ↵
      Seki N, Kodama J, Hashimoto I, et al: Thrombospondin-1 and -2 messenger RNA expression in normal and neoplastic endometrial tissues: Correlation with angiogenesis and prognosis. Int J Oncol 19:305-310, 2001
      Medline
    39. ↵
      Kononen J, Bubendorf L, Kallioniemi A, et al: Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med 4:844-847, 1998
      CrossRefMedline
    40. ↵
      Frantz GD, Pham TQ, Peale FV Jr, et al: Detection of novel gene expression in paraffin-embedded tissues by isotopic in situ hybridization in tissue microarrays. J Pathol 195:87-96, 2001
      CrossRefMedline
    41. ↵
      Jubb AM, Landon TH, Burwick J, et al: Quantitative analysis of colorectal tissue microarrays by immunofluorescence and in situ hybridization. J Pathol 200:577-588, 2003
      CrossRefMedline
    42. ↵
      Lu LH, Gillett NA: An optimized protocol for in situ hybridization using PCR-generated 33P-labeled probes. Cell Vis 1:169-176, 1994
    43. ↵
      Chalkley HW: Method for the quantitative morphologic analysis of tissues. J Natl Cancer Inst 4:47-53, 1943
    44. ↵
      Fox SB, Leek RD, Weekes MP, et al: Quantitation and prognostic value of breast cancer angiogenesis: Comparison of microvessel density, Chalkley count, and computer image analysis. J Pathol 177:275-283, 1995
      CrossRefMedline
    45. ↵
      Frank RE, Saclarides TJ, Leurgans S, et al: Tumor angiogenesis as a predictor of recurrence and survival in patients with node-negative colon cancer. Ann Surg 222:695-699, 1995
      Medline
    46. Takebayashi Y, Akiyama S, Akiba S, et al: Clinicopathologic and prognostic significance of an angiogenic factor, thymidine phosphorylase, in human colorectal carcinoma. J Natl Cancer Inst 88:1110-1117, 1996
      Abstract/FREE Full Text
    47. ↵
      Zheng S, Han MY, Xiao ZX, et al: Clinical significance of vascular endothelial growth factor expression and neovascularization in colorectal carcinoma. World J Gastroenterol 9:1227-1230, 2003
      Medline
    48. ↵
      Kim KJ, Li B, Houck K, et al: The vascular endothelial growth factor proteins: Identification of biologically relevant regions by neutralizing monoclonal antibodies. Growth Factors 7:53-64, 1992
      Medline
    49. ↵
      Presta LG, Chen H, O’Connor SJ, et al: Humanization of an anti-vascular endothelial growth factor monoclonal antibody for the therapy of solid tumors and other disorders. Cancer Res 57:4593-4599, 1997
      Abstract/FREE Full Text
    50. ↵
      Ferrara N, Hillan KJ, Gerber HP, et al: Discovery and development of bevacizumab, an anti-VEGF antibody for treating cancer. Nat Rev Drug Discov 3:391-400, 2004
      CrossRefMedline
    51. ↵
      Willett CG, Boucher Y, di Tomaso E, et al: Direct evidence that the VEGF-specific antibody bevacizumab has antivascular effects in human rectal cancer. Nat Med 10:145-147, 2004
      CrossRefMedline
    52. ↵
      Berney CR, Yang JL, Fisher RJ, et al: Vascular endothelial growth factor expression is reduced in liver metastasis from colorectal cancer and correlates with urokinase-type plasminogen activator. Anticancer Res 18:973-977, 1998
      Medline
    53. ↵
      Nanni O, Volpi A, Frassineti GL, et al: Role of biological markers in the clinical outcome of colon cancer. Br J Cancer 87:868-875, 2002
      CrossRefMedline
    54. ↵
      Cascinu S, Graziano F, Catalano V, et al: Vascular endothelial growth factor and p53 expressions in liver and abdominal metastases from colon cancer. Tumour Biol 24:77-81, 2003
      CrossRefMedline
    55. ↵
      Cascinu S, Graziano F, Catalano V, et al: Differences of vascular endothelial growth factor (VEGF) expression between liver and abdominal metastases from colon cancer: Implications for the treatment with VEGF inhibitors. Clin Exp Metastasis 18:651-655, 2000
      CrossRefMedline
    56. ↵
      Kuramochi H, Hayashi K, Uchida K, et al: Gene expression of vascular endothelial growth factor (VEGF) in primary colorectal cancer and corresponding liver metastases. Proc Amer Assoc Cancer Res 46, 2005 (abstr 5628)
    57. ↵
      Kodama J, Hashimoto I, Seki N, et al: Thrombospondin-1 and -2 messenger RNA expression in epithelial ovarian tumor. Anticancer Res 21:2983-2987, 2001
      Medline
    58. ↵
      Bossi P, Viale G, Lee AK, et al: Angiogenesis in colorectal tumors: Microvessel quantitation in adenomas and carcinomas with clinicopathological correlations. Cancer Res 55:5049-5053, 1995
      Abstract/FREE Full Text
    59. ↵
      Benjamin LE, Hemo I, Keshet E: A plasticity window for blood vessel remodelling is defined by pericyte coverage of the preformed endothelial network and is regulated by PDGF-B and VEGF. Development 125:1591-1598, 1998
      Abstract
    60. ↵
    61. DePrimo SE, Wong LM, Khatry DB, et al: Expression profiling of blood samples from an SU5416 phase III metastatic colorectal cancer clinical trial: A novel strategy for biomarker identification. BMC Cancer 3:3, 2003
      CrossRefMedline
    62. ↵
      Morimoto AM, Tan N, West K, et al: Gene expression profiling of human colon xenograft tumors following treatment with SU11248, a multitargeted tyrosine kinase inhibitor. Oncogene 23:1618-1626, 2004
      CrossRefMedline
    63. ↵
      Morgan B, Thomas AL, Drevs J, et al: Dynamic contrast-enhanced magnetic resonance imaging as a biomarker for the pharmacological response of PTK787/ZK 222584, an inhibitor of the vascular endothelial growth factor receptor tyrosine kinases, in patients with advanced colorectal cancer and liver metastases: Results from two phase I studies. J Clin Oncol 21:3955-3964, 2003
      Abstract/FREE Full Text
    64. ↵
      LeCouter J, Kowalski J, Foster J, et al: Identification of an angiogenic mitogen selective for endocrine gland endothelium. Nature 412:877-884, 2001
      CrossRefMedline
    65. ↵
      Relf M, LeJeune S, Scott PA, et al: Expression of the angiogenic factors vascular endothelial cell growth factor, acidic and basic fibroblast growth factor, tumor growth factor beta-1, platelet-derived endothelial cell growth factor, placenta growth factor, and pleiotrophin in human primary breast cancer and its relation to angiogenesis. Cancer Res 57:963-969, 1997
      Abstract/FREE Full Text
    66. ↵
      Bergers G, Javaherian K, Lo KM, et al: Effects of angiogenesis inhibitors on multistage carcinogenesis in mice. Science 284:808-812, 1999
      Abstract/FREE Full Text
    67. ↵
      Bergers G, Song S, Meyer-Morse N, et al: Benefits of targeting both pericytes and endothelial cells in the tumor vasculature with kinase inhibitors. J Clin Invest 111:1287-1295, 2003
      CrossRefMedline
    68. ↵
      Erber R, Thurnher A, Katsen AD, et al: Combined inhibition of VEGF and PDGF signaling enforces tumor vessel regression by interfering with pericyte-mediated endothelial cell survival mechanisms. Faseb J 18:338-340, 2004
      Abstract/FREE Full Text
    69. ↵
      Gerber HP, Kowalski J, Sherman D, et al: Complete inhibition of rhabdomyosarcoma xenograft growth and neovascularization requires blockade of both tumor and host vascular endothelial growth factor. Cancer Res 60:6253-6258, 2000
      Abstract/FREE Full Text
    70. ↵
      Jayson GC, Parker GJ, Mullamitha S, et al: Blockade of platelet-derived growth factor receptor-beta by CDP860, a humanized, PEGylated di-Fab′, leads to fluid accumulation and is associated with increased tumor vascularized volume. J Clin Oncol 23:973-981, 2005
      Abstract/FREE Full Text
    | Table of Contents

    This Article

    1. Published online before print December 19, 2005, doi: 10.1200/JCO.2005.01.5388 JCO vol. 24 no. 2 217-227
    1. Abstract
    2. » Full Text
    3. PDF

    Classifications

    Related Content

    1. Related Editorial

    Navigate This Article

    Current Issue

    1. July 20, 2013, 31 (21)
    1. Alert me to new issues of JCO
      • Advertisement
      • Advertisement
      • Advertisement